Thermal backfill for duct banks: compaction vs. ampacity

On a few substation and URD upgrades lately, the real bottleneck hasn’t been cable rating or relay settings. It’s duct bank and soil behavior. After the first hot summer, we logged cable temps higher than our model, and the common thread was dry-out around the encasement or a field swap to a “standard” backfill that met strength/compaction but had poor thermal performance.

Civil specs I see often center on compressive strength, slump, and density. Fair, but those choices can push us toward CLSM or crushed sand that bakes out. Our ampacity calcs assumed tested thermal backfill and moisture retention. What’s working for you long-term: flowable thermal backfill mixes, moisture-conditioned sand with fines/bentonite, geotextile wraps, or simple trench water management? Do you require thermal resistivity testing (ASTM D5334) of placed backfill, or just prequalify sources?

For those in freeze-thaw climates, any issues with settlement, heave, or fines migration when you target low thermal resistivity? What details have actually kept ampacity honest without killing constructability?